Collaborative Research: Decentralized Rules of Cohesive Swarms: From Fire Ant Rafts to Adaptive Robotic Materials

When thousands of fire ants encounter a flood, they form a dynamic, shape-shifting raft that continuously reorganizes itself to survive. The raft contracts at its core while new ants deposit at its edges via a process called treadmilling, allowing the collective to move, grow finger-like protrusions, and rapidly assemble load-bearing bridges upon contact with a solid anchor. These behaviors emerge without any central coordinator, where each ant responds only to its immediate neighbors, yet the colony collectively performs feats of adaptive engineering far beyond any single individual. Understanding how such dynamic morphing arises from simple local rules is the central scientific question of this project. Deciphering these rules, the project will generate principles transferable to robotic swarms capable of autonomous shape reconfiguration and resilient operation in unpredictable environments, with applications in flood response, deployable infrastructure, and search-and-rescue. The project will also train graduate and undergraduate students at the intersection of mechanics, biology, and robotics, and will deliver high school outreach modules on collective intelligence. All course materials, videos, and simulation tools will be made freely available to broaden access and inspire future researchers. This project addresses how cohesive active matter, exemplified by fire ant rafts, encodes adaptive, organism-level behavior through purely local interaction rules. The central hypothesis is that these behaviors obey universal principles expressible within a generalized gene-regulatory-network reaction-diffusion framework, in which morphogen-like variables couple local agent states to diffusion-driven spatial patterning. Three integrated objectives are pursued. First, high-resolution imaging and particle tracking will experimentally characterize treadmilling dynamics (such as fingering instabilities and bridge formation) with the goal of identifying key local mechanisms that drive collective morphogenesis (such as giant density fluctuations, curvature-dependent deposition, and positive feedback). Second, an agent-based model will be developed in which each agent tracks two morphogen-like state variables, an activator and an inhibitor, whose coupled evolution follows reaction-diffusion equations. This formulation will be analyzed to extract nondimensional phase diagrams linking agent-level parameters to collective states such as stable circular rafts, fingering instabilities, and directed bridge growth. Third, a swarm robotic platform of approximately fifty magnetically latching autonomous surface vehicles will be constructed and programmed with the biologically derived rules. In this system, each robot will execute the gene-regulatory-network algorithm locally to produce emergent shape morphing and treadmilling on water. Together, these efforts will establish a complete analysis, synthesis, and validation capability for engineering adaptive, self-organizing collectives. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.